OpenAI’s recently launched Dots may not be just another product in OpenAI’s confusing product matrix, encompassing ChatGPT, ChatGPT Work, and Codex. It is a powerful new user interface that allows me to work on my projects — so far still relatively small but I am getting more ambitious — much more efficiently.
Once I connect Dots to my computers and my GitHub repositories, It can do a surprising amount of work without consuming my subscription’s weekly quota. For example, I can ask it to read my blog to get the general workflow I wish for, access my local checkout on my local computer, and start working on writing code or documentation. Once everything is done, it would just commit, PR, and merge by itself. I can ask it to explain what it did over voice calls, if I wish. Calling the dot feels quite natural. I don’t get to micromanage and see the steps, but perhaps I don’t need to or even want to in the near term future. It can also go to my emails and find the ones I wanted, which is more accurate than simply entering a few keywords in the search box of my mail client.
I can see myself using the dot much more in the near-term future, and continue to explore the limits of its power.